Innovative ultrasound techniques for diagnosis of deep pelvic endometriosis: more confusion or a possible solution to the dilemma?

It was with great interest that we read the manuscript by Reid et al.1, published recently in the Journal, which proposed a new modified transvaginal ultrasound (TVS) technique (office gel sonovaginography, SVG) with the intention of improving the preoperative diagnostic accuracy of standard TVS in the assessment of deep infiltrating endometriosis (DIE) of the posterior compartment (rectosigmoid (RS), rectovaginal septum (RVS), posterior vaginal wall (PVW) and uterosacral ligaments (USL)). The authors highlighted the high accuracy of this innovative approach in predicting bowel DIE (sensitivity, 88.4%; specificity, 93.2%; positive predictive value, 79.2%; negative predictive value, 96.5%). Unfortunately, however, the accuracy of this technique in the preoperative assessment of disease involving USL, PVW and RVS sites, while maintaining high specificity (97.8%, 99.4%, 100% respectively), demonstrated a greatly reduced sensitivity (40.0%, 18.2%, 18.2%, respectively). In 2012, our study group published data regarding a ‘new technique’ known as saline contrast sonovaginography (SCSV)2, which differs from SVG solely in the contrast medium used to create an ‘acoustic window’ (using saline solution rather than ultrasound gel). Considering all pelvic DIE sites, we found SCSV to have a diagnostic accuracy, particularly specificity, comparable to that reported by Reid et al. for SVG (RS, 93.8%; USL, 95.6%; vaginal wall, 97.1%; and RVS, 100%) and in fact considerably better sensitivity (RS, 66.7%; USL, 88.9%; vaginal wall, 94.7%; RVS, 80.6%)2. Despite the fact that both these studies may appear interesting and innovative for the specialist, in our opinion neither manages to resolve the dilemma regarding which diagnostic approach would be the best method of choice for ascertaining the presence and severity of disease at each anatomical site. In the past two decades more than 50 articles have been published with the intention of defining the most accurate and reproducible sonographic technique that would be applicable on a large scale; however, we await a definitive answer and strong evidence is lacking. In a large proportion of studies, standard TVS was used as the ‘gold standard’ against which experimental techniques, such as rectal endoscopic sonography (RES) and so-called ‘modified-TVS’ techniques (including SCSV, SVG, tenderness-guided TVS, rectal-water contrast TVS, TVS with bowel preparation and TVS sliding sign), were compared. The search to identify the most effective diagnostic approach prompted Hudelist et al.3 to perform a meta-analysis of data on standard TVS, at present considered to be the first-line technique. Unfortunately, this ambitious project failed in intent in so much as it was affected by considerable bias in the data analysis and consequently in the results, since there was no discrimination between different sites of DIE in the pelvis. Considering that the different diagnostic approaches may show significant differences in accuracy related to the anatomical location of the DIE (for example, RES performed with high accuracy for intestinal endometriosis but poorly for other sites)4, we postulate that a large part of the work concerning this topic has yet to be done. Last, but not least, several studies supporting different modified-TVS techniques are perhaps affected by bias, with various claims of excellent accuracy remaining unconfirmed by other study groups or less-skilled sonographers5. The reproducibility of the results of Reid et al. and other authors proposing modified-TVS techniques in routine clinical practice is uncertain. This fact, associated with the small sample sizes and the low prevalence of single anatomical sites of pelvic DIE lesions contribute further to reduce the power of evidence reported in individual studies. In conclusion, a systematic meta-analysis is required to estimate the diagnostic power of each methodology for each specific location of DIE in the pelvis and to therefore clarify which technique should be considered the ‘first-line approach’ and which may be defined as ‘site-specific second-line’ procedures. M. Noventa†, C. Saccardi†, P. Litta†, M. Quaranta‡, D. D'Antona† and S. Gizzo*† †Department of Woman and Child Health, University of Padua, Padua, Italy; ‡Department of Obstetrics and Gynecology, University of Verona, Verona, Italy *Correspondence. (Dipartimento della Salute della Donna e del Bambino, U.O.C. di Clinica Ginecologica e Ostetrica, Via Giustiniani 3, 35128 Padova, Italy e-mail: [email protected])